This is prompted from a post by 约翰 in the iPhone discussion thread, specifically regarding the merits of a) delving deeply into the etymology of *each* character
b) studying characters more superficially - as in just the definition and being able to produce the writing

I'm personally of the opinion/experience that spending much time on etymology is counterproductive *if your goal is to learn to write characters and read native texts, soon(!)*.

I think there are a lot of interesting concepts and stories and origins and mnemonics and so on for each and every character, but all of these end up being a distraction from a primary goal of learning to understand and communicate effectively using written Chinese. Even learning a ton of radicals is of questionable utility in my opinion, as you will learn the stroke orders and common phonetics of those anyway with enough exposure to common characters, and knowing their exact *definitions* sometimes leads to more confusion, not less, with the characters containing them.

Spending a ton of time on etymology for a few characters means less time learning the plain definition and writing for many more characters, and once you are at a level where written material is accessible that serves as a great quasi-SRS system in itself - anyone that reads a article a day or a couple thousand characters from a book each night will find little benefit from using flashcards with common characters - given that, I think the immediate goal is getting to that level as fast as possible, and then using written material more, and flashcards/Skritter for those really tough/similar characters later on (and even etymology/mnemonics at that point).

Being *extremely* strict with yourself on your preferred flashcard system means you'll spend more time doing flashcards. More so if you're strict with yourself regarding remembering things like the history of the character! My own habits are to be relatively liberal with the use of the "so so" grade - I only use the "incorrect/forgotten" grade in the case that I just *completely* blanked the meaning. I still spend around 30-40 minutes per day using Skritter - and think its time well spent, but I certainly don't want to spend more time on Skritter and less on reading native material.

I just wanted to highlight this because I sometimes get the impression here that some people use Skritter (or more broadly, character acquisition) as their sole means of studying Chinese. As great as I think Skritter is, I think there is a lot of value in knowing when the returns of an additional hour of drilling down on characters are less than the returns of an hour of reading through Chinese text, even using a annotation software like Pleco's reader.

Of course, there are a group who have a strong passion for etymology, or are interested in learning the language only for a interest in characters; I don't mean to say all learners have the same goal and should pursue their goals identically. Also, I will delve into light mnemonics and etymology when I occasional encounter a *very* tough character to remember, but only as a last resort (e.g. a handful out of a couple thousand).

Personally I don't give a rats ass about etymology but I think creating mnemonics is extremely important for me. I don't care how a character evolved over the years but I do care about being able to break it apart so I can create mnemonics. I feel it isn't important what names the primitives have; it doesn't matter if I call a primitive a tree or Pikachu but it is important I give it a name so I can create mnemonics and lock the kanji into memory. I find that mnemonics greatly improve my long term memory especially when it comes to writing and not confusing similar characters for reading and writing. However I have a bad memory and I know people with good memories believe what I do is a wast of time but for me it is not.

When I was learning the first 300 characters I spent on average one hour a day with my Chinese teacher, going through New Practical Chinese Reader, half an hour going through the lesson again after work, one to two hours of vocabulary review (Pleco Chinese flashcards) and 6 hours or so of Skritter. I am spending 90% of my time doing Skritter now, trying to catch up on reviews.
I was in Beijing for two weeks and noticed that while my spoken Chinese had gotten somewhat better my characters in the wild recognition had not. That is, I could tell that I had learned a lot of the characters that I saw out there but for 30-40% of them could not for the life of me remember what they meant (I had been focussing on learning the individual characters in Skritter). Furthermore, when I then showed Skritter to my girlfriend, I could not even remember the etymology and stories of most of the characters (some of which had taken me a very long time to research). The only thing that helped to remember the characters again was to write them frequently and hear my girlfriend use them in context.
Hearing/reading/speaking/writing words in context seems to lead to a higher character acquisition rate.
As a Belgian, I had to learn 4 different languages in school and for any language there would be dictations every week. I am wondering if there are intelligible Mandarin radio shows out there? Transcribing interviews, for instance, might be a hugely beneficial learning experience.

Answered my own question by finding http://www.sinosplice.com/life/archives/2010/03/15/chinese-radio-on-the-internet.

I remember two interesting posts about taking a long time on hard stuff (intensive reading) vs. taking a short time on lots of easy stuff (extensive reading):

http://toshuo.com/2005/what-is-intensive-reading/
http://toshuo.com/2005/what-is-extensive-reading/

That recommendation was this;

"Ideally, a small portion of reading time (10%-15%) is intensive and the remainder is extensive."

It can't apply directly to one's SRS, and it's really hard to get extensive reading going in Chinese until you already know a ton of characters, but it does suggest that it's profitable to find material where you can go through a lot of words quickly, and to use a lot of extensive practice. 约翰 seems to be too far toward the intensive side of that spectrum and is not getting enough context for the characters he's learning.

Just listening to podcasts are not really this type of profitable extensive learning activity, though, since they are passive rather than active. Doing dictations from them is active, but it's very slow. Shadowing them is somewhere in between, and fast. I like shadowing:

http://learnanylanguage.wikia.com/wiki/Shadowing

I am definitely not an expert, but I've found that learning words is a lot more useful than deeply studying individual characters. For one thing, Chinese (or at least spoken Chinese) is very disyllabic. People like to put characters together rather than say them on their own. So you're not always going to hear an individual character on its own, but rather as part of a two-character word. So in that sense, the effort of knowing all of the history and meanings of an individual character is sort of wasted because you don't always encounter individual characters on their own in context. The other, related reason, is that a lot of characters, on their own, have so many different and unrelated meanings that it's really as if they have no meaning. The meaning only becomes clear when the character appears in a two-character word or a set phrase with more than one character.

Just as one example, a character like "xiang" (相) can pop up in a huge range of seemingly unrelated words -- to take a picture (照相), opposite (相反), comic dialogue (相声), each other (互相), interact (相处), similar (相似), appearance (长相) and many, many more. (And what's more, the tone changes depending on the word it's in.) I'm sure that if I studied the etymology of 相 in great depth I could see some common thread connecting all of this. But even so, I would still have to take the time to learn and memorize all of the words that contain 相. And in a listening / reading situation, knowing that someone was saying "相 + something" to me would be of very little use since that could mean practically anything.

I guess my feeling is that if you focus on the multi-character words and phrases -- i.e., what you will actually encounter in context -- then the details about individual characters and etymologies will perhaps gradually fill themselves in over time.

Interesting responses from everyone, I by no means meant to advocate my view as the end all be all, just what works for me.

@Nick - I've actually found a pair of (similar) tools which help a lot with ensuring what I'm reading falls more into the extensive margin and thus is more pleasurable as light reading material. I thought they had been discussed here before but I may be thinking of another forum... just in case:

http://www.zhtoolkit.com/posts/tools/

Both "Word List" and "Chinese Word Extractor" serve similar functionalities. They take some input text and compare it against a list of "known" words and then tell you how many *unique* words (absolute and percentage) you already do or don't know from the text, as well as list them, sortable by their frequency per X thousand words (or log-likelihood frequency), as well as by their frequency within the text being analyzed.

Basically I load up a article, extract the 3-5 most frequent unknown words and add/study them in Skritter before reading the article *the next day* using Instapaper, which works well for that sort of delayed reading (and also removes all other formatting). I've already loaded a list of my known words from Skritter so it ignores already added words and gives reports on how many unknown unique words there are in whatever is analyzed, and if I decide its too novel (e.g. more than 20-30% new) I just skip it and try another. I do the same process for books but extract around 50 new words from the analysis and study those for a week or so in Skritter before starting on the book. Of course for material where 20% of the unique words are unknown I still need to rely on an annotation/popup dictionary app like Pleco to read smoothly, but I don't let all the other unknown words slow me down - I assume if they are important enough they will eventually arise as the most-frequent-unknown in some other article down the line. Also, 20% of unknown *unique* words doesn't mean 20% of the *text* is unknown - usually it means every paragraph there are a couple of specific terms I don't know, but the rest is standard at the HSK4'ish level, and with a Pleco definition for the one or two unknowns the whole text makes sense.

The process sounds cumbersome but once you get the hang of it it only takes a couple minutes to analyze a text and add the words to Skritter, which makes reading it much easier the next day/week. I use the web-app for articles and the Windows-app for books, because longer texts tend to crash the web front, but I don't like the interface much of the app (I run it via Wine on a Mac, which has other issues).

@nick: thanks for the "Shadowing" tip.

Professor Arguelles' videos:

Shadowing Step by Step (55:39)
http://www.youtube.com/watch?v=130bOvRpt24

Shadowing Discussed (13:01)
http://www.youtube.com/watch?v=xHYDBYHi2bc

Shadowing a foreign language (Chinese) (1:08)
http://www.youtube.com/watch?v=VdheWK7u11w

And thanks to everyone else for their insightful comments as well.

The benefits of studying the history of words are usually more indirect than with other methods. It increases your awareness of which words to use and why. I love studying etymology. Most people don't. It's a personal preference.

icecream, I also love it. You learn a lot about Chinese history and culture. But the fact is that it does take a lot of time that might be better spent acquiring vocabulary through reading, listening and speaking or rote memorizing characters (with or without a more limited set of mnemonics). Extensive reading/listening/speaking could possibly help to learn the characters more swiftly. Although I find it dreadful if I do not know the true meaning of a character, my Chinese teacher keeps telling me that no child in China learns the meaning of the characters, that most adults do not know it (nor can they be bothered about it) and that I should focus on the reading/listening/speaking aspect of the language instead of spending so much time on the writing aspect. But I feel there is a bit of a chicken and the egg problem: yes, Chinese is dissyllabic, but those pairs of characters seem to consist of individual characters that enforce each other's meaning. If you do not know the meaning of the individual characters as well as the meaning of the different combinations there is no way to know which word to use in which context. So it seems fundamental to learn the sense of each and every individual character. This may also be learned through using a lot of vocabulary in many different contexts, but it seems to me that that vocabulary would have to be taught in a very structured manner (through graded readers, with accompanying audio CDs, for instance). I will find out if there is any truth to that next month.

There is a fun way to learn etymology : classical chinese. I started this year learning it in college and i find it very interesting. You have lot of funny stories and it's actually not that hard to understand.
I think classical chinese also make your understanding of chinese more solid.

The only weird part with classical is that a caractere can have so many meanings that sometimes you just don't know how to translate. sometimes with my classmates we have different stories with the same text...

I also prefer to scratch below the surface into etymology, and it's most beneficial with reading comprehension. Since I'm dipping into my first Chinese novel this month, I've found I have to guess the meanings of lots of words in context. This is exactly how reading specialists in U.S. teach word acquisition, by guessing in context before dictionary lookup.


I agree that time may be of the essence, and that quick-and-dirty memorization has situational advantages. Skritter allows me to take either approach, so thanks!

This experienced learner, who started to learn Chinese in the pre-Internet days, reflects peanutbutter's assertion that learning compound words may greatly improve character recognition and retention:

http://www.thelinguist.com/en/en/library/item/21589/

Many other interesting things as well.

The whole book seems to contain valuable nuggets of information. Providing the link here for other language learners: http://www.thelinguist.com/en/en/book-contents/.

I generally use etymology and chengyu as light relief when I'm getting bored of straight skrittering. It's interesting, and not completely wasted time in terms of learning characters. I agree that it's not the most efficient way to learn to communicate.
I think a lot of people start with characters, and then move on to words. Taking this to the next step, how valuable do you find it to work on memorizing whole phrases, or even whole sentences?